Insight 02: When Optimisation Flattened the Work

Panel III of Gathered with Care: Brittle Stars was designed to be the living panel in the series, using organic greens and browns aiming to convey the local intertidal zone of Bass Coast.

Gathered with Care: Brittle Stars, panels III, I and II. Yarra Sculpture Gallery, June 2026.

Panels I and II held their specimens relatively still, preserved and ordered within a grid. Panel III was meant to do something different. I wanted more vitality, more tonal variation, more painterly depth, with partially uncurled arms reaching toward neighbouring brittle stars so the field suggested connection rather than preservation alone. The grid would remain disciplined, each specimen still held as if collected and catalogued, but colour and gesture were intended to work against that stillness.

Variation was not decoration. It was the artistic argument of the panel.

The palette chosen to carry that argument was earthy and organic: sap greens, ochres, warm browns, tones much closer to the natural vocabulary I had worked with elsewhere in my practice. Early in the painting process that richness was visible. Some specimens were bright lime green, others sat in deeper olives or ochres, and there were enough outliers for the field to feel varied rather than uniform.

Then I started refining them.

Nothing that followed looked like a catastrophic decision. A specimen appeared slightly too bright beside its neighbours, so I muted it. Another seemed disconnected from the overall field, so it was brought closer to the dominant palette. Photographs were reviewed, layouts compared, colours adjusted, family consulted, and conversational AI provided another source of feedback during the process.

Studio iterations of the individual starfish.

The recommendations were individually reasonable. That was precisely the problem.

Across repeated iterations, the same preference kept surfacing: greater tonal consistency, fewer outliers, stronger cohesion. No single suggestion flattened the panel. Instead, each small adjustment narrowed the range slightly, and because the change was incremental it remained difficult to see while it was happening. The palette gradually moved in one direction, toward sameness.

At one stage I could see that something had drifted. The greens had moved toward a damp, dark forest register that no longer belonged to the coastal vocabulary I had intended. I went back to Kilcunda Beach and photographed sand, seaweed, shell and weathered stone as a physical reference point, then returned to the studio and brought ochres and warmer ocean-floor tones back into the palette.

Kilcunda beach rocky foreshore, photography by Dina.

That recalibration worked, but only partly.

It corrected the direction of the colour. It did not correct the deeper process already operating underneath it. I was still adjusting, balancing and resolving each specimen against the group, and every further pass continued to reduce difference.

Throughout the making, the brittle stars were laid out on tables and walls repeatedly. Their position, orientation and spacing were adjusted, the balance of the field tested, distractions reduced. There were many opportunities to look.

Yet the actual problem did not become obvious until the specimens were finished, waxed, varnished and adhered to the birch panel.

The field read flat.

That was difficult to reconcile. Visual richness is normally one of the things I trust in my own practice. The technique had worked. The individual specimens were resolved. The palette was not obviously wrong. Yet the vitality the panel had been designed around simply was not there.

I could not understand why there had been no earlier point in the process where I had seen it clearly enough to stop. Looking back at how the work had actually been made provided part of the answer.

Each brittle star had been painted horizontally, very close to the eye, under bright task lighting. Under those conditions, small differences in tone and surface treatment appeared substantial. A subtle shift from one olive to another was clearly visible. Painterly marks and variations in patination remained legible because I was looking at them from centimetres away.

Once the panel was vertical and viewed from a normal distance, the relationship changed. The waxed surface reflected light differently, and much of the tonal variation that had looked pronounced during making visually collapsed beyond roughly sixty centimetres. The differences were still physically present, but they were no longer doing the work I had assumed they were doing.

The palette had not only been optimised toward consistency. I had also been judging its success under conditions fundamentally different from those in which the artwork would eventually exist.

Vertical testing of Brittle Stars panel III.

Two things had worked against the same intention.

• One was incremental optimisation. Repeated decisions that seemed to improve cohesion gradually removed the outliers and tonal differences the panel needed.
• The other was the gap between construction conditions and final viewing conditions. Close, horizontal, brightly lit evaluation amplified differences that largely disappeared once the work was assembled and hung.

Neither problem was caused by lack of care. If anything, too much care was involved.

That has become one of the more useful things Panel III taught me. Drift does not necessarily arrive as a clearly bad artistic decision. It can emerge from many defensible decisions that accumulate toward an outcome nobody consciously chose.

Optimisation itself is not the problem.

Artists refine constantly. We edit, adjust, repaint, remove, balance and resolve because those activities are fundamental to making. The question is what those refinements are being asked to optimise.

If cohesion quietly becomes the dominant measure of success, difference may be sacrificed. If harmony is continually rewarded, the slightly awkward specimen can disappear. If every outlier is brought toward the average, a work can become technically more resolved while moving further away from its original artistic intention.

Panel III made that problem unusually visible because its intended character had been defined through variation. The panel was supposed to feel organic and alive. The qualities being progressively removed were the very ones the work had been designed to carry.

The experience also changed how I think about when an artwork is judged.

It is not enough for information to exist in the object. It has to remain perceptually available under the conditions in which the work will actually be encountered. For a wall-based work, that means testing it vertically, at distance, under changing light and with the final surface treatment operating, well before the point at which everything has been permanently assembled.

Otherwise the artist may be calibrating an experience that only exists at the workbench.

The questions that remained after Panel III were therefore more useful than simply deciding whether the artwork had succeeded.

• What exactly was the quality I had been trying to preserve?
• Which conditions were essential to carrying it, and which were secondary?
• How should those conditions be ranked when they began competing with cohesion or technical resolution?
• And how could that intention be protected throughout making rather than recognised only once drift had already occurred?

Those questions eventually contributed to a more precise understanding of what I later began calling fluency within the wider research. Not beauty as a single aesthetic ideal, but the conditions that allow a work to remain legible and coherent in the way it was intended to operate.

Panel III did not teach me to stop refining. It taught me to be more suspicious of what refinement is quietly rewarding.

The work had been optimised.

The work had also been flattened.

Completed works, Gathered with Care: Brittle Stars panel III.